Reduced 25% Inventory Waste with Demand Forecasting Model in Retail Chain
About Client
The client is a retail chain that operates across the United States with its 80+ stores offering household essentials, groceries, and fresh produce. The client serves thousands of daily customers across urban and suburban locations.
The firm managed its inventory with legacy enterprise resource planning software and spreadsheets, managed by sales professionals and store managers. However, as the SKU volume and store count grew, the traditional approach turned ineffective, causing frequent stockouts and overstocking.
Key Challenges
The manual way of inventory tracking often led to high inventory waste on perishable items, where more than 18-22% of dairy stocks and fresh produce expired before being sold. The outdated tracking system failed to detect seasonal demand swings and local preferences. Moreover, most of the stores faced frequent stockouts where 12-15% of high-velocity products used to go out of stock during the peak hours.
All of the stores used to operate on isolated demand pattern without any centralized intelligence to forecast the demand, weather shifts, regional events, and promotional impact. At the same time, a sudden spike in demand due to changed customer behavior, events, or festivals could not be captured with the legacy system.
Our Solutions
The NineHertz built an AI-powered demand forecasting solution that integrates directly into the existing ERP and POS system of the retail chain.
Smart Demand Forecasting Engine
Our group implemented a demand forecasting tool that was able to process high amounts of previous sales, weather conditions and event schedules to make 92-percent accurate predictions.
Inventory Visibility Dashboard
The new software consisted of a dashboard providing real-time data regarding the stock levels, replenishment notifications, and anticipated dates of depletion in 80 stores of the retail chain.
Automated Replenishment Recommendations
We implemented a smart solution that will automatically create order recommendations based on the current stock and supplier lead time to prevent overstock or stockouts.
Scalable Cloud-Native Architecture
The NineHertz utilized a cloud-native system which was capable of accommodating a large amount of data, new product lines and other stores without having to do any reworking or upgrade.
Impact That Drives Results
Within the 6 months of deployment of the AI-powered demand forecasting engine, the company witnessed an improvement in sales performance, operational efficiency, and inventory health.
25%
Reduced Inventory Waste
The smart forecasting system minimized the perishable write-offs of 20% to 15% in 6 months, and minimized the monthly losses in the store locations.
92%
Forecast Accuracy
The training of the model was effective to the extent that it gave a 92% accuracy to SKU level demand forecast as compared to the manual forecast which only had 50-55% accuracy.
18%
Reduced Stockouts
The automatic order recommendations and real-time access to stock levels helped eliminate the stockout instances in fast-moving SKUs by 18 per cent, regaining the lost sales and improving customer retention.
70%
Faster Procurement Cycles
The automated replenishment process saved the human workforce time by cutting the procurement planning time down to 2 hours compared to 8 hours before the automated process was implemented.
